Method for ascertaining a NOx concentration and a NH3 slip downstream from an SCR catalytic converter

Patent metadata

Field	Value
Publication number	US-11261774-B2
Application number	US-201816651104-A
Country	US
Kind code	B2
Filing date	Oct 9, 2018
Priority date	Oct 16, 2017
Publication date	Mar 1, 2022
Grant date	Mar 1, 2022

Abstract

A method is provided for ascertaining a NOx concentration and an NH3 slip downstream from an SCR catalytic converter of an internal combustion engine of a vehicle. State variables of an internal combustion engine as first input variables and an updated NH3 fill level of the SCR catalytic converter as a second input variable cooperate with at least one machine learning algorithm or at least one stochastic model. The at least one machine learning algorithm or at least one stochastic model calculates the NOx concentration and the NH3 slip downstream from the SCR catalytic converter as a function of the first input variables and the second input variables and output the same as output variables.

First claim

What is claimed is:

1. A method for ascertaining a NO x concentration and an NH 3 slip downstream from an SCR catalytic converter of an internal combustion engine of a vehicle, the method comprising the following steps:
using, by a processor, state variables of an internal combustion engine as first input variables and an updated NH 3 fill level of the SCR catalytic converter as a second input variable, for at least one machine learning algorithm or at least one stochastic model; and
calculating, by the processor via the at least one machine learning algorithm or at least one stochastic model, the NO x concentration and the NH 3 slip downstream from the SCR catalytic converter as a function of the first input variables and the second input variable;
outputting, by the processor via the at least one machine learning algorithm or the at least one stochastic model, the calculated NO x concentration and the calculated NH 3 slip downstream, as calculated output variables corresponding to an output NO x concentration and an output NH 3 slip, wherein the output NO x concentration, the output NH 3 slip, and an output NH 3 oxidation, in addition to an instantaneous NH 3 metering for the SCR catalytic converter, are input variables;
performing a stoichiometric calculation of the updated NH 3 fill level based on the input variables; and
performing, by the processor, at least one of:
controlling, as a function of the calculated output variables of the at least one machine learning algorithm or the at least one stochastic model a predictive control of an exhaust aftertreatment of the internal combustion engine or a predictive control of a drive system of the vehicle, or
establishing, as a function of the calculated output variables of the at least one machine learning algorithm or the at least one stochastic model, an exceedance of emission variables or one of outputting a corresponding warning message or initiating a corresponding error response.

2. The method as recited in claim 1 , wherein the method runs repeatedly for sequential time increments.

3. The method as recited in claim 2 , wherein the at least one machine learning algorithm or the at least one stochastic model calculates the NH 3 oxidation in the SCR catalytic converter as a function of the first input variables and the second input variable and outputs the calculated NH 3 oxidation as an output variable.

4. The method as recited in claim 3 , wherein the calculated, updated NH 3 fill level is output and is used by the at least one machine learning algorithm or the at least one stochastic model in a next time increment as the updated NH 3 fill level and as the second input variable for the calculation and output of the NO x concentration, the NH 3 slip, and the NH 3 oxidation downstream from the SCR catalytic converter.

5. The method as recited in claim 4 , wherein, in a first time increment, an initial value is selected or estimated for the updated NH 3 fill level as a function of the operating state of the internal combustion engine.

6. The method as recited in claim 4 , wherein in a first time increment, an initial NH 3 fill level of zero is selected for the updated NH 3 fill level.

7. The method as recited in claim 4 , wherein a stored initial value is selected for the updated NH 3 fill level.

8. The method as recited in claim 4 , wherein chemical reactions taking place in the SCR catalytic converter are taken into account in the stoichiometric calculation, the chemical reactions including a reduction of nitrogen oxides to nitrogen, an NH 3 oxidation, and the NH 3 slip.

9. The method as recited in claim 8 , wherein balancing equations are used for the stoichiometric calculation.

10. The method as recited in claim 1 , wherein the first input variables include at least one of: exhaust gas temperature, and/or exhaust gas pressure, and/or exhaust gas mass flow, and/or NO x concentration upstream from the SCR catalytic converter, and/or NO/NO x ratio, and/or space velocity of exhaust gas.

11. The method as recited in claim 1 , wherein the calculating takes place in the vehicle during driving operation in real-time.

12. The method as recited in claim 11 , wherein the calculating takes place in a processing unit of a control unit of the vehicle.

13. The method as recited in claim 12 , wherein the processing unit is supported in the calculations of the at least one machine learning algorithm or the at least one stochastic model by an optimized hardware unit.

14. The method as recited in claim 11 , wherein as a function of the calculated output variables of the at least one machine learning algorithm or the at least one stochastic model, monitoring or a correction of corresponding sensor output variables takes place.

15. The method as recited in claim 1 , wherein the at least one machine learning algorithm is configured as an artificial neural network.

16. The method as recited in claim 15 , wherein the at least one machine learning algorithm configured as a convolutional neural network, or a recurrent neural network, or a long short-term memory.

17. The method as recited in claim 1 , wherein the at least one stochastic model includes a Gaussian process model, or a sparse Gaussian process, or a Student-t process.

18. A non-transitory storage medium on which is stored a computer program for ascertaining a NO x concentration and an NH 3 slip downstream from an SCR catalytic converter of an internal combustion engine of a vehicle, the computer program, when executed by a computer, causing the computer to perform the following steps:
using, by a processor of the computer, state variables of an internal combustion engine as first input variables and an updated NH 3 fill level of the SCR catalytic converter as a second input variable, for at least one machine learning algorithm or at least one stochastic model; and
calculating, by the processor via the at least one machine learning algorithm or at least one stochastic model, the NO x concentration and the NH 3 slip downstream from the SCR catalytic converter as a function of the first input variables and the second input variable;
outputting, by the processor via the at least one machine learning algorithm or the at least one stochastic model, the calculated NO x concentration and the calculated NH 3 slip downstream, as calculated output variables corresponding to an output NO x concentration and an output NH 3 slip, wherein the output NO x concentration, the output NH 3 slip, and an output NH 3 oxidation, in addition to an instantaneous NH 3 metering for the SCR catalytic converter, are input variables;
performing a stoichiometric calculation of the updated NH 3 fill level based on the input variables; and
performing, by the processor, at least one of:
controlling, as a function of the calculated output variables of the at least one machine learning algorithm or the at least one stochastic model a predictive control of an exhaust aftertreatment of the internal combustion engine or a predictive control of a drive system of the vehicle, or
establishing, as a function of the calculated output variables of the at least one machine learning algorithm or the at least one stochastic model, an exceedance of emission variables or one of outputting a corresponding warning message or initiating a corresponding error response.

19. A vehicle control unit configured to for ascertaining a NO x concentration and an NH 3 slip downstream from an SCR catalytic converter of an internal combustion engine of a vehicle, t

Assignees

• Bosch Gmbh Robert

Inventors

• Daniel Christian
• Klenske Edgar
• Markert Heiner
• Schiegg Martin
• Angermaier Stefan
• Imhof Volker

Classifications

• G06N3/047

  Probabilistic or stochastic networks · CPC title

• Y02A50/20

  Air quality improvement or preservation, e.g. vehicle emission control or emission reduction by using catalytic converters · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• F01N2900/0408

  using a feed-back loop · CPC title